advanced models for simulation of cimic operations: opportunities
TRANSCRIPT
1
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
Advanced Models for Simulation ofCIMIC Operations: Opportunities & CriticalIssues provided by Intelligent Agents
Modelli Avanzati di Simulazione di Operazioni CIMIC: Criticita' eOpportunita' Offerte dagli Agenti Intelligenti
Agostino G. Bruzzone, Marina MasseiMISS DIPTEM University of Genoa{agostino,massei}@itim.unige.it - www.itim.unige.it
Alberto Tremori, Enrico BoccaMAST{alberto.tremori, enrico.bocca}@simulationteam.com
Francesca MadeoSimulation [email protected]
2
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
DIPTEM
Università
di GenovaMAST
McLeod Institute ofSimulation ScienceGenoa
DIPMEC
Unical
TURiga Universitat Autonoma
de Barcelona
MISSUniversità di Perugia
LSIS
Marseille
Universities, Research Centers and Companiesoperating worldwide in synergy for developing InnovativeSolutions with a particular focus in Modelling and Simulation
Logistica InformazioneFormazione
Euridice
AntOptima
CIREM
Università di Cagliari
CCSTra
Cagliari Australia
WhoWho WeWe Are? Are?
WindhoekNamibia
M&S NetworkMS-LES
MISS Brazil
LAMCE RIO
Blizzard srl
3
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
McLeod Institute Simulation ScienceM&S Net Genoa CenterEmail: [email protected] URL: www.itim.unige.it/mcleod
www.simulationscience.org
28 MISS 28 MISS CentersCenters, 34 M&S Net , 34 M&S Net CentersCenters World-WideWorld-Wide
The research group of DIPTEM of Genoa University is active from ‘60 inSimulation applied to Industrial Engineering and is part of MISS and M&S Net
The activities involve modeling, simulation, VV&A and analysis of IndustrialApplications and Services (design, re-engineering, management, training etc.)as: Chemical Facilities Power Plants PM
Harbor Terminals Public ServicesEnvironmentManufacturing Assembling LogisticsPublic Transportation
The Department staff is in touchworld-wide with the simulationcommunity and is presentactively to conferences, exhibitionsand working meetings with the majorAssociations, Agencies and Companies.
4
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
ObjectivesObjectives
• To identify the opportunities in simulating CIMIC by using newgeneration Computer Generated Forces (CGF) based onIntelligent Agents (IA)
• To present the critical issues and potential of Intelligent Agentsdevoted to direct new CGF
• To present the importance of Psychological & Social HumanBehavior Modifiers (HBM) in the S&R Operations
• To define a testing scenario for VV&A of this approach
• To develop Models representing the strategic issues in this framework.
5
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
Using IA, CGF and HBM inUsing IA, CGF and HBM in
Military SimulationMilitary Simulation• New Technologies and Models involving IA, CGF & HBM are a
strategic in different application areas such as:
– Training and Exercise
• Reduction of human personnel operating directly the simulation system during CAX
• New Scenarios involving dynamic simulated White Cells and Complex System (i.e.population) vs. the old pre-defined scripts
– Operational Planning
• Reducing Time for Planning Development due to the reduction of human expertsemployed in the different roles
• Possibility to Experiments different Alternatives by replicated runs carried out inAutomatic way.
– Mission rehearsal and conduct operations
• Capability to keep the simulation on-line and to conduct statistical experimentalanalysis.
6
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
Model ElementsModel Elements
for S&Rfor S&R
• Obviously in order to be successful in representing S&R(Stabilization and Reconstruction) operations from user point ofview by modeling Urban Socio-Culture-Economy Geography it iscritical to properly define the components and elements of thesimulation model to be used. In this case the model proposedherein requires an initial functionality to include:
� Civilian population entities
� Select beliefs and positions held by those entities
� Social network connecting the population entities
� Set of actions that can influence the population entities
7
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
SEGREDIFESA
CAPRICORN ProjectCAPRICORN Project
• CAPRICORN is an active EDA R&D Project devoted todevelop capabilities in the complex and critical sector ofMilitary Operation Planning, specifically for asymmetricwarfare scenarios involving CIMIC and PSYOPS, byusing CGF (Computer Generated Forces) based onIntelligent Agents (IAs)
Civil Military Co-operation And Planning Research in Complex
Operational Realistic Network
8
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
CAPRICORN benefits from the results in development of IntelligentCGF of PIOVRA (PolyfunctionalPolyfunctional Intelligent Operational Virtual Reality Intelligent Operational Virtual Reality
AgentsAgents) Project where the following Object was created:
Comportment Objects Comportment Objects are dedicated to the simulation of actors thatare dedicated to the simulation of actors that
represents behaviors of populations, movements or analog entities to where units onrepresents behaviors of populations, movements or analog entities to where units on
the field belongs.the field belongs.
Action ObjectsAction Objects units with the task of simulating elements acting the
scenario (i.e.military unit, terrorist, leader) or events (i.e.riots, demonstrations)Action Objects are referencing corresponding Comportment Object with mutualand are characterized by mutual influence
Support Objects Support Objects include objects representing influent phenomenainclude objects representing influent phenomena
corresponding to boundary conditions such as Environment Data Representationcorresponding to boundary conditions such as Environment Data Representation
including Zone, Layer and Weather Objects.including Zone, Layer and Weather Objects.
PIOVRA Background forPIOVRA Background for
CAPRICORNCAPRICORN
9
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
PIOVRA ResultsPIOVRA Results
Which are the results of PIOVRA projects to be integrated withconstructive simulation?
• A Generation of new “Intelligent” CGFs able to simulate autonomouslyhuman behavior
• Modelling of complex human behavior taking into accountpsychological and sociological parameters.
• Opportunity to model Complex Problemssuch as: CIMIC and Country Reconstruction
10
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
CIMIC Military Relevance &CIMIC Military Relevance &
CAPRICORN ProjectCAPRICORN Project• The CIMIC type activities now days constitute a significant portion
of the total military effort. The forces engaged in operations not Art.5/St.Petersburg, both for their value in the territory that for the
external visibility of the operations themselves, are under the worldopinion evaluation.
• CIMIC are critical issues in S&R (Stabilization and Normalization)operations as well as during Stabilization and Normalization phasesof most of current scenarios.
• CAPRICORN Project is focused oncreating new CGF for simulation ofCIMIC able to consider key
operational and territorial factorsalso in the planning and exercises
stages.I
11
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
Action & ComportmentAction & Comportment
ObjectsObjects• The different Units are driven by the Agents and apply different
procedures and ROE based on:
– Their configuration and Nature
– The Perception of the Scenario
– The Order Received
– The Previous Experiences
12
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
Example of Mutual InfluenceExample of Mutual Influence
among Ethnicsamong Ethnics
The Attitudes among the IA are based on Fuzzy Representation and their interactionare regulated y Artificial Intelligence techniques
13
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
IA-CGF IA-CGF MODULESMODULES
The new IA-CGF Modules devoted to create the CIMICsimulation include:
•IA-CGF Units
•IA-CGF Human Behaviors
•IA-CGF Non-Conventional Frameworks
14
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
IA-CGF IA-CGF UnitsUnitsIA-CGF Units are a set of interoperable units with capability tobe integrated in constructive simulation
– Police
– Gangs
– Local Population
– Rioters
– Insurgents
– Terrorist
– Local Authorities
– Warlord
– Criminal Organizations
– NGOs (CIMIC ops.)
– Civil Personnel (CIMIC ops.)
– Domestic/National Situation (for instance for troops moral):• Population• Media• Lobbies
– International Public Opinion
– International Diplomacy
– New Threats (i.e. 2nd Generation Terrorists)
These are examples of non-conventional units
controlled by IA-CGF
15
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
IA-CGF Human BehaviorsIA-CGF Human Behaviors
Specific modules with IA-CGF Human Behaviors:
• Fear
• Stress
• Fatigue
• Training Level
• Aggressiveness
• Ethnic Factors
• Religious Factors
• Combat Skills/Experience
IA-CGF Human Behaviors operate as a set of furthercharacteristics to be added to each unit in constructivesimulation.i.e. now in constructive simulation every unit in the scenario have infos about status andtype of ammo, by IA-CGF it will be added dynamic information about level of fear andstress and the Units performing according to it
16
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
IA-CGF Non-ConventionalIA-CGF Non-Conventional
FrameworksFrameworksIt is important to consider the integration in a scenario of the IA-CGF-Non-Conventional Frameworks (IA-CGF-NCF), eachsimulating specific events:
•IA-CGF CIMIC/HUMANITARIAN FRAMEWORKS– Food Distribution
– Reconstruction
•IA-CGF Homeland Security and Civil Protection FRAMEWORKS– Natural Disaster
– Evacuation
•IA-CGF PSYOPS and INTELLIGENCE FRAMEWORKS– Possible integration with Sibilla© Serious Game for Intelligence Officers
training
In non conventional scenarios for particular training purposes.We can imagine to have active different non conventional Frameworks,in different locations, with different level of detail inside thesimulated theater.
17
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
CIMIC Comportment ObjectsCIMIC Comportment Objects
• In the CIMIC scenario Comportment Objects will reproduce:
– Population � Ethnic Groups� Cultural Layers� Social Layers� Local Communities
– Public Institutions� Governmental Entities� Local Administrations� Schools� Health Care Systems� Political Parties
� Military Organizations
� Local Army
� Allied Forces
� Opposite Forces
� Neutral Forces
� Paramilitary Organizations
� Police Forces
� Warlords
� Insurgents
� Terrorists
18
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
Interoperable ModelsInteroperable Models
The evolution of the
situation requires models
such as:
Event
ZoneObjects
ZoneObjects
ZoneObjects
ZoneObject
ComportmentObject
Action Object
Action Object
Attenuation Factors
Media
TV
Radio
Press
Web
Action Object
Impact Factors
DirectPerception
Media Diffusion
Action Object
Action Object
Action Object
Action Object
Social Interactions
Mutual
Influence
Action
Comportment
– Civilian Attitude Model: describes
changes in population attitude as a
result of Host Nation actions and
insurgency activities, and perceived
good/bad actions.
– Social Network Model: helps to
describe civilian response to new
factors that interact with civilian
behaviors, social structures and
specific communication parameters
as hierarchy and message contents.
– Economic Model: describes how the
economic decision of a Host Nation
can affect and impact the attitude and
activities of entities.
19
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
VV&A FeaturesVV&A FeaturesFeature/Objective Present Example and Note1. User-Defined Initializing Parameters YES The user defined the profile of the Gang as well as the ROE to be
used by Blue Units
2. Analyze Surrounding Envinroment and React RespectivelyCapability
YES The Blue Units encountering the Riot and the Gang takes actions tostop the looting
3. Cooperation Capacity YES Some Blue unit are providing support on others reaching thedemonstration/riot
4. Force Aggregating/Disaggregating Capability and relevantmilitary hierarchy
YES Disaggregation of Blue unit in two Squads after dissolution of the Riot
5. Resultant Aggregation Levels different fromaggregating/disaggregating elements sum/subtraction
YES The combination of Demonstration and Gang looting create impact onthe area different from the sum of the single entities and introducesthe generation of a riot
6. Limit Proper Autonomy to Achieve Common ObjectiveCapability
YES It is possible to enable/disable the possibility for the Blue Unit torequest direct support to the other ones and to let the scenario evolvewith this other condition
7. Stress Level Indicator applicable for the entities behaviordefinition
YES These aspects affect both Population and Military Units all along thesimulation
8. Implementation of Typical Human Behavior (survival instinctand moral/ethical motivations)
YES It is possible to enable/disable the feature and check, versus criticalriots, the different respect of ROE by Military Units
9. Distinct Friend, Foe and Neutral Units YES Distinction betweeen Gang and Militia
10. Explicit ROE justifying Proper Behavior YES Each entity provides a Log including the conditions under what eachdifferent ROE applied
11. Military Reports to Higher Commanders Capability YES Reporting that includes encounters with other units, Riots andsituation evolution
12. Decision Process Traceability YES Each entity provides a Log related to the factors affecting their actions
13. Feedback capability YES Blue Unit moving among cells of an ethnic group affects thepopulation evolution and the eventual creation of a Riot
14. CGF simulated - single entity YES A single entity is representing the agitators that change the attitude ofthe demontration/riot
15. CGF simulated - team YES The militia unit is corresponding to a team
16. CGF simulated - squad YES The Blue Unit in patrol corresponds to a Squad
17. CGF simulated - platoon YES The Blue Unit providing support corresponds to a Platoon
18. HLA Integration YES Reports about actions and events are distributed as interaction in theHLA Federation during Simulation Runs
Example of Testing
the Features in PIOVRA
It is critical to guarantee proper VV&A for models especially dueto HBM and Interoperability Issues
20
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
Example of ScenarioExample of Scenario
Experimental AnalysisExperimental Analysis
Example of Response Surface Methodology and Design of Experimentsobtained by running PIOVRA scenarios and correlating urban stabilizationmetrics with operational planner actions
21
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
ConclusionsConclusions
• Authors’ aims is to develop a architectures, models and algorithms path for
creating common framework for supporting S&R foundations. Main
objective is to model CIMIC in SASO (Stability and Support Operations)
scenarios.
• The major benefit expected by applying IA-CGF in CIMIC operations is the
possibility to introduce many agents able to drive and direct actions of
involved entities and organizations with strong and complex interactions; so
that
• The Intelligent Agents make their decisions
on scenario awareness and mutual interactions
dynamically changing
22
DIPTEMUniversità di Genova
MAST© 2010 Copyright MISS DIPTEM
LSIS
Marseille
www.simulationteam.com
DIPTEM University of Genoavia Opera Pia 1516145 Genova, Italywww.itim.unige.itAgostino G. [email protected]
DIPTEM
MASTSimulation Lab, via Molinero 117100 Savona, Italywww.mastsrl.euAlberto [email protected]
MAST
ReferencesReferences
LSIS
LSISAv Escadrille Normandie Niemen13397 Marseilles, Francewww.lsis.orgClaudia [email protected]